JOURNAL OF PROPULSION AND POWER Vol. 26, No. 3, May–June 2010
Simultaneous Pressure and Optical Measurements of Nanoaluminum Thermites: Investigating the Reaction Mechanism K. Sullivan∗ and M. R. Zachariah† University of Maryland, College Park, Maryland 20740 DOI: 10.2514/1.45834 This work investigates the reaction mechanism of metastable intermolecular composites by collecting simultaneous pressure and optical signals during combustion in a constant-volume pressure cell. Nanoaluminum and three different oxidizers are studied: CuO, SnO2 , and Fe2 O3 . In addition, these mixtures are blended with varying amounts of WO3 as a means to perturb the gas release in the system. The mixtures with CuO and SnO2 exhibit pressure signals that peak on timescales faster than the optical signal, whereas the mixtures containing Fe2 O3 do not show this behavior. The burn time is found to be relatively constant for both CuO and SnO2 , even when a large amount of WO3 is added. For Fe2 O3 , the burn time decreases as WO3 is added, and the temperature increases. The results are consistent with the idea that oxidizers such as CuO and SnO2 decompose and release gaseous oxidizers fast, relative to the burning, and this is experimentally seen by an initial pressure rise followed by a prolonged optical emission. In this case, the burning is rate limited by the aluminum, and it is speculated to be similar to the burning of aluminum in a pressurized oxygenated environment. For the Fe2 O3 system, the pressure and optical signals occur concurrently, indicating that the oxidizer decomposition is the rate-limiting step.
the fuel and oxidizer, and energy propagation. None of these phenomena themselves are well understood for nanoparticles, especially when the heating rate is high, as is the case in combusting systems. Nanoaluminum has been shown to have a much lower ignition temperature than micron-sized aluminum. Although both have a naturally formed oxide shell surrounding the elemental core, in the nanoparticle, the oxide shell can account for a relatively large portion of the particle’s mass. Upon heating, the aluminum core melts at a much lower temperature than the oxide shell (933 vs 2327 K) and can expand, inducing stresses on the oxide shell. The response of the shell to this expansion may be different for a nanoparticle vs a large particle, leading to a lower ignition temperature. Some authors argue that a decomposition or phase change in the shell occurs, thus allowing aluminum to diffuse outward [21–23], whereas other authors argue that the rapid expansion of the core induces enough stress to completely shatter the shell and unload the aluminum as small liquid clusters [24–26]. The burning mechanism of aluminum thereafter will be quite different, depending on what mechanism of ignition happens. The burning mechanism of nanoaluminum particles is currently poorly understood. For combustion-type applications, the heating rate of nanoparticles will be high (106 –108 K=s). Experiments should be designed to reproduce these heating rates, and one such experimental technique that accomplishes this is a shock tube. Bazyn et al. [27,28] studied the combustion of nanoaluminum at elevated temperatures and pressures in a shock tube. The authors combust aluminum at varying temperatures, pressures, and oxygen mole fractions, and they use three-color pyrometry to measure the particle temperature. The authors show that the ambient temperature plays a significant roll on the aluminum combustion, indicating that heat losses are much more important for nanoparticles than for largersized particles. The same authors [29] show that a transition from a diffusion to a kinetic-limited mechanism begins to occur below a critical particle size, 10 m. For a kinetic-limited mechanism, the flame sits closer to (if not on) the particle surface, and the flame temperature is limited by the boiling point of aluminum. The third phenomena occurring in the reaction mechanism of a self-propagating MIC is energy propagation, and authors [12,30] have shown that the dominant mode of energy propagation through a loose powder is convection. As a result, MICs often exhibit an optimal reactivity that correlates with gas production instead of temperature. For example, Sanders et al. [12] found that Al=CuO has
I. Introduction
M
ETASTABLE intermolecular composites (MICs) are a class of energetic materials consisting of an intimate mixture of fuel and oxide nanoparticles. Aluminum is primarily used as the fuel, and a variety of metal oxides have been used, including but not limited to CuO, WO3 , MoO3 , Bi2 O3 , and Fe2 O3 . MICs are a relatively new class of energetic materials, and research efforts to understand them have increased since Aumann et al. [1] reported a 1000x increase in reactivity when nano-sized Al=MoO3 particles were used in place of their micron-sized counterparts. The high energy density and wide range of tunability of MICs make them attractive candidates for uses in propellants, pyrotechnics, and explosives. However, the reaction mechanism is still very poorly understood. A commonly used technique to prepare MICs is to ultrasonicate the powders in a dispersing liquid, such as hexane or isopropyl alcohol, and then allow the liquid to dry. The remaining powder can be broken up or sieved until it has the consistency of a loose powder. A variety of experimental methods have been used to investigate the reactivity of these powders, including thermal analysis [2,3], combustion in a shock tube [4], flame propagation in open channels [5–11] and tubes [12–14], heated filament studies [3], and constantvolume pressure cells [6,15–19]. The pressure signal and/or optical emission can be collected to investigate the reactivity of these materials. The pressurization rate has been shown to correlate with flame propagation velocities [20] and is typically reported as a relative measurement of reactivity. Other authors [12,14] have shown a correlation between the peak pressure and the propagation velocity. Recently, authors [12–14] have used an instrumented burn tube to collect the optical and pressure signals simultaneously. If the reaction is self-propagating, there are three phenomena occurring simultaneously: ignition of the material, reaction between Received 4 June 2009; revision received 11 January 2010; accepted for publication 22 January 2010. Copyright © 2010 by Michael R. Zachariah. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission. Copies of this paper may be made for personal or internal use, on condition that the copier pay the $10.00 per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923; include the code 0748-4658/10 and $10.00 in correspondence with the CCC. ∗ Department of Mechanical Engineering. † Department of Mechanical Engineering and Department of Chemistry and Biochemistry;
[email protected]. 467
468
SULLIVAN AND ZACHARIAH
a peak reactivity for an equivalence ratio very near stoichiometric. The authors use equilibrium calculations to show that a stoichiometric mixture produces the maximum amount of Cu gas, and any deviation from this mixture will lower the temperature, hindering the gas production, and thus the convective mode of energy propagation. Conversely, other mixtures often exhibit enhanced reactivity for slightly fuel-rich mixtures. The same authors show that an Al=Bi2 O3 thermite has a greater propagation velocity and peak pressure for an equivalence ratio of 1.3 when compared with an equivalence ratio of 1.0, even though the calculated adiabatic flame temperature is a few hundred degrees lower at the fuel-rich condition. Also, Al=MoO3 shows an optimal reactivity for an equivalence ratio around 1.2–1.4. The enhancement is attributed to increased gas production for the fuel-rich conditions and is predicted by thermodynamic equilibrium calculations. Both Sanders et al. [12] and Malchi et al. [14] show that the peak pressure correlates with the flame propagation velocity. In the two works, an instrumented burn tube is used to simultaneously collect the pressure and optical signals. The authors use equilibrium calculations to show correlations between the predicted equilibrium gas and the experimental trends in pressure. From Fig. 9 in Malchi et al. [14], it appears that the optical signal reaches its peak on the same time scale as the pressure does, 10 s.
II.
III. Experimental The aluminum used in this study was purchased from the Argonide Corporation and is designated as 50 nm ALEX by the supplier. The aluminum was found to be 70% active by mass, as measured by thermogravimetric analysis. All other materials were purchased from Sigma Aldrich and have average particle diameters of less than 100 nm, as specified by the supplier. All samples were prepared by weighing out the powder and adding it to a ceramic crucible. Approximately 10 ml of hexane was added, and the mixtures were ultrasonicated in a sonicating bath for 30 min to ensure intimate mixing. The samples were then placed in a fume hood until the hexane evaporated and the wetness was gone, and then the samples were put in a 100 C furnace for a few minutes to drive off any remaining hexane. The dry powders were very gently broken up with a Teflon-coated spatula until the consistency was that of a loose powder. Appropriate equipment (antistatic mats and wrist straps, along with basic lab safety equipment, such as gloves and goggles) should be used when handling the dried powder in order to minimize the risk of accidental ignition and injury. A fixed mass (25 mg) of the powder was weighed out and placed in a small (13 cc free volume) combustion cell. The sample sits in a bowl-shaped sample holder, and a Nichrome coil contacts the top of the powder so the reaction propagates downward and into the holder upon ignition. Two ports (located on the sides of the cell) were used to collect the pressure and optical signal simultaneously. In one port, a lens tube assembly, containing a plano–convex lens (f 50 mm), collected light and imaged onto an optical fiber coupled to a highspeed Si photo detector (1 ns rise time, model DET10A, Thorlabs). In the second port, a piezoelectric pressure sensor was employed, the details for which can be found in Prakash et al. [17]. The powder is ignited by manually increasing the voltage and current until the sample is ignited. This is done as rapidly as possible to avoid significant heating of the powder. The data collection was triggered by the rising optical signal. There is always an 60 s delay between the onset of the optical emission and the onset of the pressure signal. This is due to the time delay between the optical triggering and when the pressure wave arrives at the sensor, a few centimeters away. The pressure data were thus shifted in time for the analysis so that the onset of the pressure and light are shown to occur simultaneously.
Thermochemistry of Mixtures
Recent mass spectrometry work by our group has indicated that oxygen release from the metal oxide decomposition is important in the reaction mechanism of thermites, in particular for CuO and Fe2 O3 . The current work expands on this idea to investigate the burning of nanoaluminum composites in a constant-volume pressure cell. The pressure and optical signals are collected simultaneously to have two different measurements of reactivity. The oxides studied are CuO, Fe2 O3 , and SnO2 . These particular oxidizers have adiabatic flame temperatures at or above the boiling point of the metal in the metal oxide, and the gas is predicted to be almost entirely composed of this metal at equilibrium. These oxidizers also decompose to suboxides and gaseous oxidizers, which will be discussed in more detail later. The calculated equilibrium for stoichiometric mixtures of these oxidizers with aluminum is shown in Table 1. The CHEETAH 4.0 code was used with the JCZS product library [31], as recommended by Sanders et al. [12]. The mixture density was assumed to be 0:00192 g=cc, because we always react 25 mg of material in our 13 cc cell. The experimental pressurization rate is also given for comparison. We will start by investigating the simultaneous pressure and optical signals for the three oxidizers mentioned previously. We will then go on to perturb the system by adding increasing amounts of WO3 in place of the metal oxide. We chose WO3 because, when added as the minor component, the adiabatic temperature remains relatively unchanged. Also, WO3 is predicted to produce very little equilibrium gas and does not decompose to O2 or any significant gaseous oxidizing species until greater than 2800 K. All blends are stoichiometric and are referred to in terms of the molar percentage of WO3 in the oxidizer. For example, a 40% WO3 mixture means that 40% of the oxidizer molecules are WO3 , 60% are the other oxidizer, and the corresponding amount of aluminum is added to make the overall mixture stoichiometric, assuming complete conversion to Al2 O3 .
Table 1
Results and Discussion
We first show the simultaneous pressure and optical signals for pure Al/CuO, Al=SnO2 , and Al=Fe2 O3 in Fig. 1. Also included is pure Al=WO3 for comparison. Note that the axes for each plot have all been adjusted to fill the plot area. From Fig. 1, we can immediately see that CuO and SnO2 exhibit a pressure peak well before the optical signal reaches its peak. In the case of Fe2 O3 and WO3 , the pressure and optical signals occur concurrently. It is important to take a moment to discuss our interpretation of the optical signal and the various considerations that may complicate the analysis. First of all, an accurate measurement of the temperature for such a large sample is greatly complicated by the fact that the viewing area is optically thick, and thus the measurement would be biased to the outermost (or coolest) region of the reaction. Also, we have no reason to believe that the flame region would be spatially homogeneous. It is possible that the optical signal could be measuring the emission from large chunks of material that ignite later in time;
Calculated temperature and gas production for stoichiometric mixtures of various metal oxides with nanoaluminum
Metal oxide Boiling point metal, K CuO SnO2 Fe2 O3 WO3
IV.
2837 2533 3023 5933
Tad (Cheetah UV), K
Moles gas/kg reactant
Contribution of metal to the total gas, %
Experimental pressurization rate, psi/s
2967 2573 2834 3447
3.5 2.2 0.52 0.13
97 94 98
